A huge part of the chip crisis is automotive demand. Mainly that automotive cars demand older chips that the semi companies don't want to invest the capital to make (increased wafer size). Essentially there's a "cycle mismatch". This has been a long time coming though...

A few years ago I was talking to someone at Toyota. He was concerned that the amount of "code" in a car was growing exponentially. I forget the exact dates, but relatively recently, it was about 200 lines of code...and now it's aroudn 200,000 (bad metric, but still).

Semi manufacturers move at a different cycle. Huge capital investments in fabs means that the fabs have moved to a greater density wafers and have no interest in restarting their older assets to serve automotive.

Automotive manufacturers have 7 year design cycles...so while they *do* update the design before the car is out of production, it's no trivial task to upgrade the electronics.

By the time the automotive companies finish designing a car, the chips are out of date.

No easy policy solution here! "I would simply invest in a new fab" is Twitter VC bullshit. Any new fab will ALSO produce newer chips!

The only hope here is automotive design cycles continue to shorten. I can see competitive pressures moving in this direction (tesla etc).